Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
YC News Traffic So Far (ycombinator.com)
20 points by pg on March 9, 2007 | hide | past | favorite | 32 comments


The RSS feed could do with links back to this comments page - Google reader doesn't show any such link.

That would increase traffic from RSS users, I should think.


concur, this would match my usage pattern much better (read interesting articles from RSS, then bring up another window to read the comments)


Here's what gets me- in the first couple of days there were some pretty good posts that got _tons_ of marks. There are some users in the top 20 rankings that haven't had any activity since the first couple of days- like http://news.ycombinator.com/user?id=beau . I have to be honest, while the customer service article was nice, I don't think it would get more than 20 points these days- I don't think any article would.

So I thought there must have been a lot of people using it at the very beginning- excited and contributing and the karma flowed like milk and honey and then the novelty wore off and a bunch of people dropped out. I see the opposite in the graph. I guess the usage of the site is what changed considerably.


The reason those early articles got so many points is that the decay function was broken initially. Instead of accelerating downward off the frontpage with age, the velocity of stories' fall actually decreased. Because of this bug some of those stories stayed on the frontpage racking up votes for 4 days.


Looking at old comments[1], though, indicates that the overall volume of karma flow was much higher in the early days. Of course, it could be that the bug-driven hyperkarma on posts led users to be more generous with karma on comments.

[1] In particular, the feature request thread.


It might be interesting to try a decay function on the leader board.

Now, it's probably not cool to take away karma points, but it would reward activity. Plus, it'll get around the issue of single posters who post a single ridiculously popular post. They would have their hayday in karma but then slide back in the rankings in favor of the regulars.


The karma of active users would continue to rise whilst the dead users stay steady and eventually get overtaken.

Perhaps a 'most karma in the past week' leader board would work?


I like the idea; this mirrors some discussion in the "Thoughts on using news.YC" thread, viz. http://news.ycombinator.com/comments?id=2987


There are a lot of people who follow the links yet dont vote even if they like them. An example is the facebook YC news group. It has 42 members yet the story announcing it only has 13 points.

Could this be an example of the 1% rule? http://technology.guardian.co.uk/weekly/story/0,,1823959,00.html


I tend to think that it is more the result of a bit of cognitive dissonance, particularly with respect to what the points values mean. They can be viewed either as a metric of popularity, or a metric of quality. Most people {{handwaving}} conceive of points as some mixture of the two, hence there is a reluctance to give a point to a story with, say, 20 points, even if you liked it, because you don't think it is a "21-point quality" story.


I think one of the drops could have been when RSS was implemented and us junkies didnt have to do constant page refreshes to keep up to date.


I used the RSS feed for a little while, but I tend to just leave a browser window open at yc.news now as I work on various things.

There's nothing quite like the YC tan and orange to really put me in the startup mood.


Err, that's page views, not unique IPs? If so, then I think I'm about half of that graph...

Edit: Nevermind, I'm about a 1/20th. I slipped a decimal.


Come to think about it there is a drop when I went away for a few days...


Interesting...could someone who's been around longer than I have take a stab at assigning some of those peaks to real-world events? I'm particularly interested in the peak near the end of February, and the following trough.


I think it was an O'Reilly blog post about news.yc.


pg, did you create news.yc as a long term solution or is this something temporary (just for the startup school and YC applicants)? It's great to see the site's enormous growth. Unfortunately, there is a good chance the traffic will drop sharply following April.

*ill still be around, though. YC can't get rid of me that easily*


Long term, certainly. There are a lot more features coming.


We got another day's stats so as long as the story was still live I just updated the url. Here's the old one: http://ycombinator.com/ycnewstraffic.png


Do you mind if I ask how many users there are? Have memberships mirrored the page requests more or less?


Oddly enough I'd never checked before, but as of now there are 1426 accounts. The reddits told me once that about 1 out of 10 visitors there is logged in, so this seems about right.


Do we have any metrics on liveness of accounts? I've certainly felt that the community of active, thoughtful commentators is pretty small (not a bad thing!).


Based on the stats of other web apps I would predict that there are around 140 active users who comment and submit articles. This is however just a guess.


That's a surprisingly low number, actually. The first SFP had 227 applications[1] and I assume it's grown since then. Based on nothing at all, I'll guess that the average application has 2.35 founders, so that's 533 applicants. YC has been pretty explicit that they want applicants to participate here.

[1] http://web.archive.org/web/20050414040411/ycombinator.com/sfpannounce.html


Well, here's one:

lynx -source http://news.ycombinator.com/user?id=$USER_IN_QUESTION | grep '<span class="comhead">' | wc -l


It works, but only under the assumption that user liveness is constant or increasing over time. A liveness metric should differentiate strongly between a user who has commented 30 times in the past 4 days, and someone who commented 25 times in the first 4 days of news.yc. and 5 times since. Perhaps I'll hack something together.


This site is so much better than Reddit these days. I don't even check R anymore (too much noise/politics/pics/videos).

Thanks PG!


What tool do you use to analyse traffic?


wc -l


Why don't you go further and count unique ip's? cut -f2 | sort -u | wc -l (Assuming ip is the second field after timestamp in your logs)


Sorry to sound stupid. What is wc -l? What does it stand for?


Prints the number of lines in a file.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: